Abstract

STUDY OBJECTIVES--To recruit a representative sample of farmworkers, accurately quantify the range and extent of their animal exposures, and measure the associated risks of illness. DESIGN--Inception cohort. SETTING--The study was undertaken among farmworkers living in five local authority areas in the catchment of Hereford and Preston Public Health Laboratories, England. PARTICIPANTS--A quota sample of 404 people on 255 agricultural holdings took part. The holdings were selected at random from the Ministry of Agriculture, Fisheries and Food register. Altogether 58% of eligible subjects approached agreed to participate. MEASUREMENTS AND MAIN RESULTS--The sample had the same sex distribution as the 1991 census for those giving their occupation as agriculture. The mean age was significantly (p < 0.01) higher (44.6 years v 42.2 years) than that of those giving their occupation as agriculture, forestry or fishing in the census, although the modal range (45-59 years) was the same. At enrollment interviews, subjects individually reported contact with up to nine animal species (mode 4) out of 26 reported in all. Based on the numbers contacted and the frequency and intimacy of contact, scores on a ranked ordinal scale from 0-5 were constructed for each species and frequencies for each score were plotted. Subjects also reported past operations and serious illness. A history of pneumonia was significantly (p < 0.05) associated with a pigeon loft on the farm (relative risk (RR) 7.3) and attending farrowing pigs (RR 6.6), and one of leptospirosis with a rat problem on the farm (RR 28.1). Cattle contact was associated with a significantly lower likelihood (protective) of glandular fever (RR 0.19) and rheumatic or scarlet fever (RR 0.12). These effects were significantly related to rankings of the extent of exposure. CONCLUSIONS--It is possible to recruit a representative sample of farmworkers and measure their animal exposures in great detail. Among these exposures, associations with plausible risk factors for pneumonia and leptospirosis and apparently protective factors for glandular fever, scarlet fever, and rheumatic fever have been demonstrated, which further show a relationship between the extent of exposure and response. These findings can be tested further by examining the relationship of exposures to serological evidence of illness or by further prospective follow up of this and similarly well characterised cohorts, or both.